An observational mechanism for detection of distributed denialof-service attacks

Abstract

This study proposes a continuous mechanism for detecting distributed denial of service (DDoS) attacks from network traffic data. The mechanism aims to systematically organise traffic data and prepare them for DDoS attack detection using convolutional deep-learning neural networks. The proposed mechanism contains ten phases covering activities, including data preprocessing, feature selection, data labelling, model building, model evaluation, DDoS detection, attack pattern identification, alert creation, notification delivery, and periodical data sampling. The evaluation results suggested that the detection model built based on convolutional deep-learning neural networks and relevant network traffic features provided 97.2% detection accuracy. The study designed a holistic mechanism that considers the systematic network traffic data management for continuous monitoring and good performance of DDoS attack detection. The proposed mechanism could provide a solution for network traffic data management and enhance the existing methods for DDoS attack detection. In addition, it generally contributes to the cybersecurity body of knowledge.